- Email: [email protected]
The potential of the baobab (Adansonia digitata L.) as a proxy climate archive
Applied Geochemistry Applied Geochemistry 21 (2006) 1674–1680 www.elsevier.com/locate/apgeochem
The potential of the baobab (Adansonia digitata L.) as a proxy climate archive I. Robertson
a,*
, N.J. Loader a, C.A. Froyd b, N. Zambatis c, I. Whyte c, S. Woodborne d
a Department of Geography, University of Wales Swansea, Swansea SA2 8PP, UK Centre for the Environment, University of Oxford, South Parks Road, Oxford OX1 3QY, UK c Scientific Services, Kruger National Park, Private Bag X402, Skukuza 1350, South Africa Quaternary Dating Research Unit, CSIR Environmentek, P.O. Box 395, Pretoria 0001, South Africa b
d
Available online 1 September 2006
Abstract The large girth and immense size of the baobab has caused many to speculate about its age. Unfortunately reliable age estimates cannot be determined from growth rates as the girth varies in response to different moisture regimes. In a similar way, ages cannot be determined from ring-width measurements or X-ray densitometry as the absorbent nature of the soft fibrous wood and distortion upon drying prevent the application of these techniques. The Southern Hemisphere bomb radiocarbon curve was used to demonstrate that the rings of a recently-fallen baobab (Adansonia digitata L.) from Kruger National Park appear to be annual. The detrended C isotope values of finely-ground wholewood from another baobab specimen were found to be highly associated with January precipitation (r = 0.72; p < 0.01). This study demonstrates that high resolution information about past climates may be obtained by analysing the C isotope values from baobab samples even if distortion of ring-widths has occurred during drying. However, this relationship must be replicated before the baobab can be demonstrated to be a reliable palaeoclimatic proxy. Ó 2006 Elsevier Ltd. All rights reserved.
1. Introduction The African baobab (Adansonia digitata L.) is perhaps one of the most recognisable of all African tree species. It is distributed across the majority of Africa, south of the Sahara, yet our knowledge of it is surprisingly limited (Wickens, 1982). The trees can grow to immense size and this had led to considerable speculation about the age of specimens.
*
Corresponding author. E-mail address: [email protected] (I. Robertson).
In AD 1749, the French naturalist Michel Adanson examined two living trees on the Magdalene Islands off Cape Verde on the coast of Senegal, which he estimated to be 5150 years old (Wickens, 1982). However, the African missionary and explorer, David Livingstone, questioned these calculations as a tree of this age would pre-date the Great Flood of the Old Testament. Assuming annual growth rings, Livingstone calculated that a baobab 85 feet (26 m) in circumference was 14 centuries old and consequently within the Christian era (Livingstone, 1861). Even today, estimates of the oldest tree ages vary and some of the less than scientific
0883-2927/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.apgeochem.2006.07.005
I. Robertson et al. / Applied Geochemistry 21 (2006) 1674–1680
age estimates have been perpetuated through the literature. Although Wilson (1988) stated that few baobabs exceed 400 years; published maximum age estimates range from 2 ka (Coates Palgrave and Coates Palgrave, 2002) to 3 ka (Esterhuyse et al., 2001) with some samples reported to be over 4 ka old (van Wyk, 2000). Irregular and asymmetrical growth (Wickens, 1982), seasonal changes in girth caused by water losses and gains (Fenner, 1980) and the shrinkage of older trees make it extremely difficult to measure the girth of baobabs accurately and caution is required if growth rates are to be used to determine age (Weyerhaeuser, 1985; Guy, 1982). If the rings are annual, as Livingstone assumed, the cross-dating of samples (Stokes and Smiley, 1963) would enable the establishment of chronologies with absolute confidence in the integrity of dates. However, although the soft, fibrous wood is arranged in concentric layers (Baum, 1995), ‘‘Practically every authority consulted states categorically that baobabs do not form annual rings.’’ (Guy, 1970). Although the absorbent nature of the wood structure may preclude the use of the baobab for conventional dendroclimatology or X-ray densitometry, the potential exists for stable isotope techniques to be applied to this archive as a measure of past climates. As baobabs are known to be very sensitive to climate (Guy, 1970), research is required to expand upon the pioneering work of Swart (1963) to determine if the rings are annual. The aim of this pilot research project is to determine if the rings of the baobab are annual and, if so, can any proxy parameters be related to climatic variables.
1675
(24°59.21 0 S, 31°34.92 0 E; circumference = 1.5 m). Both of these samples were obtained in April 2002. In the laboratory, the samples were air dried over several weeks to minimise the effects of shrinkage. Although the fragile nature of the segment prevented the use of a measuring stage, ring-widths were measured using a hand lens and graduated scale. To examine the nature of the concentric rings, 3 samples obtained from the segment of the recently-fallen baobab (Tree 1) were selected for high-precision 14C analysis at the Quaternary Dating Research Unit, Pretoria, South Africa. The samples (ca. 30 g) were collected from identifiable growth bands and the number of growth bands between each sample counted. Each sample was then split into matchstick size pieces and pre-treated using standard techniques to remove soluble organic components (Tans and Mook, 1980). The 14 C content of the samples was determined by gas proportional counting (Vogel and Marais, 1971). Additionally, samples from each ring of the core (22 in total -Tree 2) were removed as a fine homogeneous powder using an electric drill, taking care not to char or fractionate the sample of wood. Stable C isotope measurements were determined using a PDZ Europa 20/20 continuous flow stable isotope ratio mass spectrometer and ANCA GSL elemental analyser. Results are expressed using the conventional notation as deviations from the Vienna Pee Dee belemnite (VPDB) standard (Coplen, 1995). The average precision on replicate results was 0.12&, which reflects sample preparation and measurement errors. 3. Results
2. Methods The locations of recently-fallen baobabs were recorded during the annual aerial wildlife census in Kruger National Park in August AD 2001. It is thought that many of these trees died sometime after the extreme floods that occurred during January and February AD 2000. A radial segment was obtained from the base of a recently fallen tree close to the Luvuvhu River near to Pafuri (22°25.09 0 S, 31°14.18 0 E; circumference = 13.3 m) (Tree 1). The tree was hollow at the base with some roots present. There were several branches with leaves indicating that some sections of the tree were still actively photosynthesising. An additional 12 mm diameter core was obtained from a living tree of known age (Tree 2) that was planted in Skukuza in AD 1972
The atmospheric testing of nuclear weapons caused the Northern Hemisphere atmospheric 14C levels to peak at almost double the natural value in AD 1963–1964. Even though the majority of tests were carried out in the Northern Hemisphere, the Southern Hemisphere experienced a similar change albeit of a reduced magnitude in AD 1965 (Vogel et al., 2002; Hua et al., 2003). As the elevated 14C value of atmospheric CO2 is still above natural levels and trees directly record atmospheric CO2 (Hua et al., 1999), Southern Hemisphere tree-ring samples can be dated to approximately the nearest year from AD 1955 to present (Vogel et al., 2002). Although the technique is relatively straightforward, a 14C value will usually have two possible calibrated dates depending upon the intercept with the ‘bomb
1676
I. Robertson et al. / Applied Geochemistry 21 (2006) 1674–1680
carbon’ curve (Vogel et al., 2002). Initial calibration with the Southern Hemisphere 14C dataset revealed two equally valid calibrated dates for each sample from Tree 1 (Table 1). However, using a priori information about the time-series, such as that the inner portion of the segment must be relatively older than the outer parts, it was possible to eliminate improbable dates. Of the 38 rings in the segment, the three samples for radiocarbon dating were obtained from rings 6, 9–10 and 26, where ring 1 is located towards the centre of the section (Fig. 1). The outer sample (ring 26) had a 14C value of 157.9 ± 0.4 pMC (PTA-8789) which corresponds to a calibrated radiocarbon date of AD 1959 or AD 1989/90 (Table 1). The date of the middle sample (rings 9–10) which had a 14C value of 148.6 ± 0.6 pMC (PTA-8731) must be older than the outer sample (Table 1). Whatever
date is chosen for the middle sample (AD 1963 or AD 1972), it is clear that AD 1959 can now be rejected for the outer sample. By a similar process of elimination, it is possible to reject the date of AD 1963 for the middle sample (rings 9–10). Therefore, the gap of 16–17 rings between the outer (ring 26) and the middle sample (rings 9–10) corresponds to the period of 17–18 years (Fig. 1). Within the errors of radiocarbon dating, the rings of the recently fallen baobab (Tree 1) would appear to be essentially annual in nature. By sampling at the extreme limits of a plant’s distribution, it is likely that the plant response to climate will be enhanced relative to extraneous noise (Schweingruber, 1988). To test this hypothesis, a 12 mm diameter core was obtained from a baobab (Tree 2) growing at the Staff Housing at Skukuza (24°59.21 0 S, 31°34.92 0 E) to enable the association
Table 1 Calibration of radiocarbon dates using independently-dated plant and animal material and direct Southern Hemisphere atmospheric measurements from the year AD 1959 (Vogel et al., 2002) Sample reference
Sample designation
d13CVPDB (&)
14
Calibrated Southern Hemisphere date (AD)
PTA-8789 PTA-8731 PTA-8739
Ring 6 (inner) Rings 9–10 (middle) Ring 26 (outer)
25.1 25.6 26.1
157.9 ± 0.4 148.6 ± 0.6 117.3 ± 0.5
1964 or 1968/9 1963 or 1972 1959 or 1989/90
C (pMC)
The high precision radiocarbon values of Tree 1 were determined at the CSIR in Pretoria, South Africa (Vogel and Marais, 1971). pMC denotes the radiocarbon content as percent modern carbon relative to the standard reference sample (NBS Oxalic acid).
180
Ring 6: 1964/5 or 1968/69 170
Rings 9-10: 1972 160
14
C( pMC)
150 140
Ring 26: 1989/90
130 120 110 100 90 1950
1955
1960
1965
1970
1975
1980
1985
1990
1995
2000
Year (AD) Fig. 1. The average annual summer 14C values (pMC) for mid-southern latitudes (Vogel et al., 2002) from AD 1950 to 2000. Superimposed upon this figure are the 14C values for the samples obtained from rings 6, 9–10 and 26 of the recently-fallen baobab (Tree 1). The error bars on the radiocarbon dates represent the standard 1r errors.
I. Robertson et al. / Applied Geochemistry 21 (2006) 1674–1680
with temperature and precipitation recorded at the nearby Skukuza meteorological station (24°59 0 S, 31°36 0 E) to be investigated. The mean monthly precipitation at Skukuza is illustrated for the period 1981–2002 (Fig. 2). It is evident that the majority of precipitation falls during the summer months (October–April). The stable C isotope values were measured on the 22 apparent rings from the core sample (Fig. 3). Based upon the results from the baobab segment (Tree 1), it was assumed that baobabs could exhibit annual growth rings. Although some studies assign the calendar year to the year in which growth commenced (LaMarche et al., 1979), the calendar year was assigned to the period of maximum growth and therefore, it was estimated that the core sample covered the period AD 1981–2002. For this short modern time series, the raw C isotope values (Fig. 3) were detrended to remove possible nonclimatic trends (i.e.: changing atmospheric 13C composition) by fitting a linear trend through the data. One possible limitation of this technique of statistical detrending is that for longer time series potential low-medium frequency climatic trends could also be removed (McCarroll and Loader, 2004). The relationship between detrended C isotope values and meteorological variables was investigated using SPSS/WIN. Over the period AD 1981–2002, the
1677
highest association was found between January precipitation and the detrended C isotope series (r = 0.72; p < 0.01) (Fig. 4). 4. Discussion: annual rings and potential for palaeoclimate research Stomatal conductance and photosynthetic rate are the two key factors that determine the C isotope value of trees rings (McCarroll and Loader, 2004). Baobabs have the ability to minimise water loss by decreasing stomatal conductance (Fenner, 1980). However, the positive relationship between detrended C isotope values and precipitation cannot be explained adequately by the widely reported equations for C3 plants: d13 Cp d13 Ca a ðb aÞðci =ca Þ where d13Cp and d13Ca represent the C isotope values of the plant and atmosphere respectively; a represents the diffusional fractionation factor; b is the biochemical fractionation and ci and ca represent the intercellular and ambient concentrations of CO2 (Vogel, 1980; Farquhar et al., 1982). Although most studies refer to the baobab as a tree; it is strictly a succulent with a poorly determined ecology or physiology (Fenner, 1980) and with the exception of Adansonia digitata, the genus
120
Precipitation (mm)
100
80
60
40
20
0 Jan
Feb
Mar
Apr
May
Jun
Jly
Aug
Spt
Oct
Nov
Dec
Month Fig. 2. Mean monthly precipitation recorded at the Skukuza meteorological station (24°59 0 S, 31°36 0 E; elevation 263 m) over the period 1981–2002.
1678
I. Robertson et al. / Applied Geochemistry 21 (2006) 1674–1680
-25.0
13
CVPD B (‰)
-26.0
-27.0
-28.0
-29.0 1981
1986
1 991
1996
2001
Year (AD)
350
1.5
300
1.0
250 0.5 200 0.0 150 -0.5 100 -1.0
50
0 1981
Detrended carbon isotopes (‰)
January precipitation (mm)
Fig. 3. Carbon isotope values determined on finely ground wholewood from Tree 2 at Skukuza over the period 1981–2002. The error bars represent the precision of replicate measurements (typically ± 0.12&).
-1.5 1986
1991
1996
2001
Year (AD) Fig. 4. Relationship between January precipitation (open circles) and the detrended C isotope series (closed circles) from Tree 2 at Skukuza (r = 0.72; p < 0.01).
is poorly known (Baum, 1995). However, it is known that baobabs have the ability to survive extremely dry winters by storing summer moisture in their swollen trunks and consequently baobab wood can contain a high proportion of water (Kelly,
2000). When leaf photosynthetic activity resumes, changes in stomatal conductance or photosynthetic rate must be invoked to explain the strong positive relationship of January precipitation with detrended C isotope values (Fig. 4). Growth-limiting factors
I. Robertson et al. / Applied Geochemistry 21 (2006) 1674–1680
are typically reported as being responsible for the isotopic fractionation observed for plants. Consequently, if the rate controlling an environmental variable varies at either side of the optimum response of plant growth, a positive or negative relationship with that variable could be observed (Schleser et al., 1999). Whatever the actual mechanism of the observed response, it is clear that the relationship between January precipitation and the detrended C isotope series (r = 0.72) is significant at the 99% confidence limit (Fig. 4) and demonstrates the potential of C isotope values of baobabs as a palaeoclimatic proxy. The fragile nature of baobabs, together with possible shrinkage upon drying limit the effectiveness of conventional dendroclimatology or X-ray densitometry to obtain meaningful information about past climates. However, this pilot study has demonstrated that such information can be obtained by determining C isotope values of baobab wood even if distortion of ring-width has occurred during drying. Palaeoclimate records from tropical regions are essential to our understanding past changes in the Earth’s climate system, equator-pole linkages and the sensitivity of tropical regions to future climate change. In an attempt to increase the number of high resolution palaeoclimatic records from Africa, ring-width chronologies have been carefully constructed from species previously thought impossible to date (Stahle et al., 1999; Couralet et al., 2005) and radiocarbon dating has been used to confirm tentative time-series (Vogel et al., 2001; Robertson and Woodborne, 2002). The three radiocarbon dates from the radial segment from Tree 1 demonstrate that the Adansonia digitata specimen growing in Kruger National Park appears to form annual rings (Fig. 1). Although it is tempting to state that baobabs have annual rings and hence, the baobab has the potential as a palaeoclimate proxy covering millennia, replication of these initial results is required to confirm these potentially far-reaching conclusions. Analysis of baobab tissue using radiometric and stable C isotope analysis would suggest that this is indeed the case, however, the nature of the climate in this region is also such that false/partial absent rings may also be a characteristic of this species. Such a growth response is to be found in many tree species growing in marginal regions worldwide where growth is routinely influenced by external controls. Whilst this makes conventional dendro-
1679
chronology difficult, objective construction of absolutely-dated chronologies is still possible. Conventional analyses of ring width variability may likely fail with this particular species, owing to its propensity for shrinkage and deformation; however, assuming high levels of series replication within and between plants, development of isotopic chronologies may instead provide future potential for chronology building and the study of environmental change since the isotopes are not subject to these limitations. Swart (1963) reported a baobab with an estimated age of 1010 ± 10 years, consequently, if these results can be replicated across its distribution, there is significant potential to further explore carefully dated, pristine baobabs as long term palaeoclimatic proxies with high temporal resolution. Acknowledgements The authors thank numerous members of staff at Kruger National Park for assistance. In particular, we are grateful to Michele Hofmeyr, Bruce Leslie, Harry Biggs (Kruger National Park) and Bob Scholes (CSIR) for constructive advice and assistance; Andre Potgeiter (Kruger National Park) for planting the tree at Skukuza!; Molly McKnight (Oregon) and Million Cossa (Kruger National Park) for help with fieldwork; Gill Collet (CSIR) and Paula Stanlake (University of Wales Swansea) for technical support; and John Vogel (CSIR) and Glenda Swart (South African Weather Service) for the provision of unpublished data. Permission to carry out research was granted by the Conservation Services Management Committee at Kruger National Park. The authors (IR, CAF and NJL) thank the European Union (grants EVK2-CT-2002-00136 and EVK2-CT-2002-00147) and NERC (NE/B501504/ 1) for financial assistance. References Baum, D.A., 1995. A systematic revision of Adansonia (Bombacaceae). Ann. Mo. Bot. Garden 82, 440–471. Coates Palgrave, M., Coates Palgrave, K., 2002. Trees of Southern Africa, third ed. Struik Publishers (Pty) Ltd., Cape Town. Coplen, T., 1995. The discontinuance of SMOW and PDB. Nature 375, 285. Couralet, C., Saas-Klaassen, U., Sterck, F., Bekele, T., Zuidema, P.A., 2005. Combining dendrochronology and matrix modelling in demographic studies: an evaluation for Juniperus procera in Ethiopia. Forest Ecol. Manage. 216, 317–330.
1680
I. Robertson et al. / Applied Geochemistry 21 (2006) 1674–1680
Esterhuyse, N., von Breitenbach, J., So¨hnge, H., 2001. Remarkable Trees of South Africa. Briza Publications, Pretoria. Farquhar, G.D., O’Leary, M.H., Berry, J.A., 1982. On the relationship between carbon isotope discrimination and the intercellular carbon dioxide concentration in leaves. Aust. J. Plant Physiol. 9, 121–137. Fenner, M., 1980. Some measurements on the water relations of baobab trees. Biotropica 12, 205–209. Guy, G.L., 1970. Adansonia digitata and its rate of growth in relation to rainfall in south Central AfricaProceed Transaction, vol. 54. Rhodesia Scientific Association, Salisbury, pp. 68–84. Guy, G.L., 1982. Baobabs and elephants. Afr. J. Ecol. 20, 215– 220. Hua, .Q., Barbetti, M., Worbes, M., Head, J., Levchenko, V.A., 1999. Review of radiocarbon data from atmospheric and tree ring samples for the period 1945–1997 AD. IAWA J. 20, 261– 283. Hua, Q., Barbetti, M., Zoppi, U., Chapman, D.M., Thomson, B., 2003. Bomb radiocarbon calibration from northern New South Wales, Australia: implications for dendrochronology, atmospheric transport, and air-sea exchange of CO2. Radiocarbon 45, 431–447. Kelly, H.L.P., 2000. The effect of elephant utilisation of the Sterculia rogersii and Adansonia digitata populations of the Kruger National Park. Unpublished MSc Thesis, University of Pretoria. LaMarche Jr., V.C., Holmes, R.L., Dunwiddie, P.W., Drew, L.G., 1979. Tree-Ring Chronologies of the Southern Hemisphere: 5. South Africa. Chronology Series V, Laboratory of Tree-Ring Research, University of Arizona. Livingstone, D., 1861. A Popular Account of Missionary Travels and Researches in South Africa. John Murray, London. McCarroll, D., Loader, N.J., 2004. Stable isotopes in tree rings. Quatern. Sci. Rev. 23, 771–801. Robertson, I., Woodborne, S.W., 2002. Carbon isotopes confirm the competitive advantages of Prosopis over Acacia erioloba.
In: Study of Environmental Change using Isotope Techniques, IAEA-CSP-13/P. International Atomic Energy Agency, Vienna, pp. 33–434. Schleser, G.H., Helle, G., Lu¨cke, A., Vos, H., 1999. Isotope signals as climate proxies: the role of transfer functions in the study of terrestrial archives. Quatern. Sci. Rev. 18, 927–943. Schweingruber, F.H., 1988. Tree Rings: Basics and Applications of Dendrochronology. D. Reidel, Boston. Stahle, D.W., Mushove, P.T., Cleveland, M.K., Roig, F., Haynes, G.A., 1999. Management implications of annual growth rings in Pterocarpus angolensis from Zimbabwe. Forest Ecol. Manage 124, 217–229. Stokes, M.A., Smiley, T.L., 1963. An Introduction to Tree-Ring Dating. University of Chicago Press, Chicago. Swart, E.R., 1963. The age of the baobab tree. Nature 373, 134– 137. Tans, P.P., Mook, W.G., 1980. Past atmospheric CO2 levels and 13 C/12C ratios in tree rings. Tellus 32, 268–283. van Wyk, P., 2000. Field Guide to the Trees of the Kruger National Park. Struik Publishers (Pty) Ltd., Cape Town. Vogel, J.C., 1980. Fractionation of the carbon isotopes during photosynthesis. In: Sitzungsberichte der Heidelberger Akademie der Wissenschaften. Springer-Verlag, Berlin, pp. 111–135. Vogel, J.C., Marais, M., 1971. Pretoria radiocarbon dates I. Radiocarbon 13, 378–394. Vogel, J.C., Fuls, A., Visser, A., 2001. Radiocarbon adjustments to the dendrochronology of a yellowwood tree. S. Afr. J. Sci. 97, 164–166. Vogel, J.C., Fuls, A., Visser, A., 2002. Accurate dating with radiocarbon from the atom bombtests. S. Afr. J. Sci. 98, 437– 438. Weyerhaeuser, F.J., 1985. Survey of elephant damage to baobabs in Tanzania’s Lake Manyara National Park. Afr. J. Ecol. 23, 235–243. Wickens, G.E., 1982. The Baobab Africa’s upside down tree. Kew Bull. 37, 173–209. Wilson, R.T., 1988. Vital statistics of the baobab (Adansonia digitata). Afr. J. Ecol. 26, 197–206.
The potential of the baobab (Adansonia digitata L.) as a proxy climate archive I. Robertson
a,*
, N.J. Loader a, C.A. Froyd b, N. Zambatis c, I. Whyte c, S. Woodborne d
a Department of Geography, University of Wales Swansea, Swansea SA2 8PP, UK Centre for the Environment, University of Oxford, South Parks Road, Oxford OX1 3QY, UK c Scientific Services, Kruger National Park, Private Bag X402, Skukuza 1350, South Africa Quaternary Dating Research Unit, CSIR Environmentek, P.O. Box 395, Pretoria 0001, South Africa b
d
Available online 1 September 2006
Abstract The large girth and immense size of the baobab has caused many to speculate about its age. Unfortunately reliable age estimates cannot be determined from growth rates as the girth varies in response to different moisture regimes. In a similar way, ages cannot be determined from ring-width measurements or X-ray densitometry as the absorbent nature of the soft fibrous wood and distortion upon drying prevent the application of these techniques. The Southern Hemisphere bomb radiocarbon curve was used to demonstrate that the rings of a recently-fallen baobab (Adansonia digitata L.) from Kruger National Park appear to be annual. The detrended C isotope values of finely-ground wholewood from another baobab specimen were found to be highly associated with January precipitation (r = 0.72; p < 0.01). This study demonstrates that high resolution information about past climates may be obtained by analysing the C isotope values from baobab samples even if distortion of ring-widths has occurred during drying. However, this relationship must be replicated before the baobab can be demonstrated to be a reliable palaeoclimatic proxy. Ó 2006 Elsevier Ltd. All rights reserved.
1. Introduction The African baobab (Adansonia digitata L.) is perhaps one of the most recognisable of all African tree species. It is distributed across the majority of Africa, south of the Sahara, yet our knowledge of it is surprisingly limited (Wickens, 1982). The trees can grow to immense size and this had led to considerable speculation about the age of specimens.
*
Corresponding author. E-mail address: [email protected] (I. Robertson).
In AD 1749, the French naturalist Michel Adanson examined two living trees on the Magdalene Islands off Cape Verde on the coast of Senegal, which he estimated to be 5150 years old (Wickens, 1982). However, the African missionary and explorer, David Livingstone, questioned these calculations as a tree of this age would pre-date the Great Flood of the Old Testament. Assuming annual growth rings, Livingstone calculated that a baobab 85 feet (26 m) in circumference was 14 centuries old and consequently within the Christian era (Livingstone, 1861). Even today, estimates of the oldest tree ages vary and some of the less than scientific
0883-2927/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.apgeochem.2006.07.005
I. Robertson et al. / Applied Geochemistry 21 (2006) 1674–1680
age estimates have been perpetuated through the literature. Although Wilson (1988) stated that few baobabs exceed 400 years; published maximum age estimates range from 2 ka (Coates Palgrave and Coates Palgrave, 2002) to 3 ka (Esterhuyse et al., 2001) with some samples reported to be over 4 ka old (van Wyk, 2000). Irregular and asymmetrical growth (Wickens, 1982), seasonal changes in girth caused by water losses and gains (Fenner, 1980) and the shrinkage of older trees make it extremely difficult to measure the girth of baobabs accurately and caution is required if growth rates are to be used to determine age (Weyerhaeuser, 1985; Guy, 1982). If the rings are annual, as Livingstone assumed, the cross-dating of samples (Stokes and Smiley, 1963) would enable the establishment of chronologies with absolute confidence in the integrity of dates. However, although the soft, fibrous wood is arranged in concentric layers (Baum, 1995), ‘‘Practically every authority consulted states categorically that baobabs do not form annual rings.’’ (Guy, 1970). Although the absorbent nature of the wood structure may preclude the use of the baobab for conventional dendroclimatology or X-ray densitometry, the potential exists for stable isotope techniques to be applied to this archive as a measure of past climates. As baobabs are known to be very sensitive to climate (Guy, 1970), research is required to expand upon the pioneering work of Swart (1963) to determine if the rings are annual. The aim of this pilot research project is to determine if the rings of the baobab are annual and, if so, can any proxy parameters be related to climatic variables.
1675
(24°59.21 0 S, 31°34.92 0 E; circumference = 1.5 m). Both of these samples were obtained in April 2002. In the laboratory, the samples were air dried over several weeks to minimise the effects of shrinkage. Although the fragile nature of the segment prevented the use of a measuring stage, ring-widths were measured using a hand lens and graduated scale. To examine the nature of the concentric rings, 3 samples obtained from the segment of the recently-fallen baobab (Tree 1) were selected for high-precision 14C analysis at the Quaternary Dating Research Unit, Pretoria, South Africa. The samples (ca. 30 g) were collected from identifiable growth bands and the number of growth bands between each sample counted. Each sample was then split into matchstick size pieces and pre-treated using standard techniques to remove soluble organic components (Tans and Mook, 1980). The 14 C content of the samples was determined by gas proportional counting (Vogel and Marais, 1971). Additionally, samples from each ring of the core (22 in total -Tree 2) were removed as a fine homogeneous powder using an electric drill, taking care not to char or fractionate the sample of wood. Stable C isotope measurements were determined using a PDZ Europa 20/20 continuous flow stable isotope ratio mass spectrometer and ANCA GSL elemental analyser. Results are expressed using the conventional notation as deviations from the Vienna Pee Dee belemnite (VPDB) standard (Coplen, 1995). The average precision on replicate results was 0.12&, which reflects sample preparation and measurement errors. 3. Results
2. Methods The locations of recently-fallen baobabs were recorded during the annual aerial wildlife census in Kruger National Park in August AD 2001. It is thought that many of these trees died sometime after the extreme floods that occurred during January and February AD 2000. A radial segment was obtained from the base of a recently fallen tree close to the Luvuvhu River near to Pafuri (22°25.09 0 S, 31°14.18 0 E; circumference = 13.3 m) (Tree 1). The tree was hollow at the base with some roots present. There were several branches with leaves indicating that some sections of the tree were still actively photosynthesising. An additional 12 mm diameter core was obtained from a living tree of known age (Tree 2) that was planted in Skukuza in AD 1972
The atmospheric testing of nuclear weapons caused the Northern Hemisphere atmospheric 14C levels to peak at almost double the natural value in AD 1963–1964. Even though the majority of tests were carried out in the Northern Hemisphere, the Southern Hemisphere experienced a similar change albeit of a reduced magnitude in AD 1965 (Vogel et al., 2002; Hua et al., 2003). As the elevated 14C value of atmospheric CO2 is still above natural levels and trees directly record atmospheric CO2 (Hua et al., 1999), Southern Hemisphere tree-ring samples can be dated to approximately the nearest year from AD 1955 to present (Vogel et al., 2002). Although the technique is relatively straightforward, a 14C value will usually have two possible calibrated dates depending upon the intercept with the ‘bomb
1676
I. Robertson et al. / Applied Geochemistry 21 (2006) 1674–1680
carbon’ curve (Vogel et al., 2002). Initial calibration with the Southern Hemisphere 14C dataset revealed two equally valid calibrated dates for each sample from Tree 1 (Table 1). However, using a priori information about the time-series, such as that the inner portion of the segment must be relatively older than the outer parts, it was possible to eliminate improbable dates. Of the 38 rings in the segment, the three samples for radiocarbon dating were obtained from rings 6, 9–10 and 26, where ring 1 is located towards the centre of the section (Fig. 1). The outer sample (ring 26) had a 14C value of 157.9 ± 0.4 pMC (PTA-8789) which corresponds to a calibrated radiocarbon date of AD 1959 or AD 1989/90 (Table 1). The date of the middle sample (rings 9–10) which had a 14C value of 148.6 ± 0.6 pMC (PTA-8731) must be older than the outer sample (Table 1). Whatever
date is chosen for the middle sample (AD 1963 or AD 1972), it is clear that AD 1959 can now be rejected for the outer sample. By a similar process of elimination, it is possible to reject the date of AD 1963 for the middle sample (rings 9–10). Therefore, the gap of 16–17 rings between the outer (ring 26) and the middle sample (rings 9–10) corresponds to the period of 17–18 years (Fig. 1). Within the errors of radiocarbon dating, the rings of the recently fallen baobab (Tree 1) would appear to be essentially annual in nature. By sampling at the extreme limits of a plant’s distribution, it is likely that the plant response to climate will be enhanced relative to extraneous noise (Schweingruber, 1988). To test this hypothesis, a 12 mm diameter core was obtained from a baobab (Tree 2) growing at the Staff Housing at Skukuza (24°59.21 0 S, 31°34.92 0 E) to enable the association
Table 1 Calibration of radiocarbon dates using independently-dated plant and animal material and direct Southern Hemisphere atmospheric measurements from the year AD 1959 (Vogel et al., 2002) Sample reference
Sample designation
d13CVPDB (&)
14
Calibrated Southern Hemisphere date (AD)
PTA-8789 PTA-8731 PTA-8739
Ring 6 (inner) Rings 9–10 (middle) Ring 26 (outer)
25.1 25.6 26.1
157.9 ± 0.4 148.6 ± 0.6 117.3 ± 0.5
1964 or 1968/9 1963 or 1972 1959 or 1989/90
C (pMC)
The high precision radiocarbon values of Tree 1 were determined at the CSIR in Pretoria, South Africa (Vogel and Marais, 1971). pMC denotes the radiocarbon content as percent modern carbon relative to the standard reference sample (NBS Oxalic acid).
180
Ring 6: 1964/5 or 1968/69 170
Rings 9-10: 1972 160
14
C( pMC)
150 140
Ring 26: 1989/90
130 120 110 100 90 1950
1955
1960
1965
1970
1975
1980
1985
1990
1995
2000
Year (AD) Fig. 1. The average annual summer 14C values (pMC) for mid-southern latitudes (Vogel et al., 2002) from AD 1950 to 2000. Superimposed upon this figure are the 14C values for the samples obtained from rings 6, 9–10 and 26 of the recently-fallen baobab (Tree 1). The error bars on the radiocarbon dates represent the standard 1r errors.
I. Robertson et al. / Applied Geochemistry 21 (2006) 1674–1680
with temperature and precipitation recorded at the nearby Skukuza meteorological station (24°59 0 S, 31°36 0 E) to be investigated. The mean monthly precipitation at Skukuza is illustrated for the period 1981–2002 (Fig. 2). It is evident that the majority of precipitation falls during the summer months (October–April). The stable C isotope values were measured on the 22 apparent rings from the core sample (Fig. 3). Based upon the results from the baobab segment (Tree 1), it was assumed that baobabs could exhibit annual growth rings. Although some studies assign the calendar year to the year in which growth commenced (LaMarche et al., 1979), the calendar year was assigned to the period of maximum growth and therefore, it was estimated that the core sample covered the period AD 1981–2002. For this short modern time series, the raw C isotope values (Fig. 3) were detrended to remove possible nonclimatic trends (i.e.: changing atmospheric 13C composition) by fitting a linear trend through the data. One possible limitation of this technique of statistical detrending is that for longer time series potential low-medium frequency climatic trends could also be removed (McCarroll and Loader, 2004). The relationship between detrended C isotope values and meteorological variables was investigated using SPSS/WIN. Over the period AD 1981–2002, the
1677
highest association was found between January precipitation and the detrended C isotope series (r = 0.72; p < 0.01) (Fig. 4). 4. Discussion: annual rings and potential for palaeoclimate research Stomatal conductance and photosynthetic rate are the two key factors that determine the C isotope value of trees rings (McCarroll and Loader, 2004). Baobabs have the ability to minimise water loss by decreasing stomatal conductance (Fenner, 1980). However, the positive relationship between detrended C isotope values and precipitation cannot be explained adequately by the widely reported equations for C3 plants: d13 Cp d13 Ca a ðb aÞðci =ca Þ where d13Cp and d13Ca represent the C isotope values of the plant and atmosphere respectively; a represents the diffusional fractionation factor; b is the biochemical fractionation and ci and ca represent the intercellular and ambient concentrations of CO2 (Vogel, 1980; Farquhar et al., 1982). Although most studies refer to the baobab as a tree; it is strictly a succulent with a poorly determined ecology or physiology (Fenner, 1980) and with the exception of Adansonia digitata, the genus
120
Precipitation (mm)
100
80
60
40
20
0 Jan
Feb
Mar
Apr
May
Jun
Jly
Aug
Spt
Oct
Nov
Dec
Month Fig. 2. Mean monthly precipitation recorded at the Skukuza meteorological station (24°59 0 S, 31°36 0 E; elevation 263 m) over the period 1981–2002.
1678
I. Robertson et al. / Applied Geochemistry 21 (2006) 1674–1680
-25.0
13
CVPD B (‰)
-26.0
-27.0
-28.0
-29.0 1981
1986
1 991
1996
2001
Year (AD)
350
1.5
300
1.0
250 0.5 200 0.0 150 -0.5 100 -1.0
50
0 1981
Detrended carbon isotopes (‰)
January precipitation (mm)
Fig. 3. Carbon isotope values determined on finely ground wholewood from Tree 2 at Skukuza over the period 1981–2002. The error bars represent the precision of replicate measurements (typically ± 0.12&).
-1.5 1986
1991
1996
2001
Year (AD) Fig. 4. Relationship between January precipitation (open circles) and the detrended C isotope series (closed circles) from Tree 2 at Skukuza (r = 0.72; p < 0.01).
is poorly known (Baum, 1995). However, it is known that baobabs have the ability to survive extremely dry winters by storing summer moisture in their swollen trunks and consequently baobab wood can contain a high proportion of water (Kelly,
2000). When leaf photosynthetic activity resumes, changes in stomatal conductance or photosynthetic rate must be invoked to explain the strong positive relationship of January precipitation with detrended C isotope values (Fig. 4). Growth-limiting factors
I. Robertson et al. / Applied Geochemistry 21 (2006) 1674–1680
are typically reported as being responsible for the isotopic fractionation observed for plants. Consequently, if the rate controlling an environmental variable varies at either side of the optimum response of plant growth, a positive or negative relationship with that variable could be observed (Schleser et al., 1999). Whatever the actual mechanism of the observed response, it is clear that the relationship between January precipitation and the detrended C isotope series (r = 0.72) is significant at the 99% confidence limit (Fig. 4) and demonstrates the potential of C isotope values of baobabs as a palaeoclimatic proxy. The fragile nature of baobabs, together with possible shrinkage upon drying limit the effectiveness of conventional dendroclimatology or X-ray densitometry to obtain meaningful information about past climates. However, this pilot study has demonstrated that such information can be obtained by determining C isotope values of baobab wood even if distortion of ring-width has occurred during drying. Palaeoclimate records from tropical regions are essential to our understanding past changes in the Earth’s climate system, equator-pole linkages and the sensitivity of tropical regions to future climate change. In an attempt to increase the number of high resolution palaeoclimatic records from Africa, ring-width chronologies have been carefully constructed from species previously thought impossible to date (Stahle et al., 1999; Couralet et al., 2005) and radiocarbon dating has been used to confirm tentative time-series (Vogel et al., 2001; Robertson and Woodborne, 2002). The three radiocarbon dates from the radial segment from Tree 1 demonstrate that the Adansonia digitata specimen growing in Kruger National Park appears to form annual rings (Fig. 1). Although it is tempting to state that baobabs have annual rings and hence, the baobab has the potential as a palaeoclimate proxy covering millennia, replication of these initial results is required to confirm these potentially far-reaching conclusions. Analysis of baobab tissue using radiometric and stable C isotope analysis would suggest that this is indeed the case, however, the nature of the climate in this region is also such that false/partial absent rings may also be a characteristic of this species. Such a growth response is to be found in many tree species growing in marginal regions worldwide where growth is routinely influenced by external controls. Whilst this makes conventional dendro-
1679
chronology difficult, objective construction of absolutely-dated chronologies is still possible. Conventional analyses of ring width variability may likely fail with this particular species, owing to its propensity for shrinkage and deformation; however, assuming high levels of series replication within and between plants, development of isotopic chronologies may instead provide future potential for chronology building and the study of environmental change since the isotopes are not subject to these limitations. Swart (1963) reported a baobab with an estimated age of 1010 ± 10 years, consequently, if these results can be replicated across its distribution, there is significant potential to further explore carefully dated, pristine baobabs as long term palaeoclimatic proxies with high temporal resolution. Acknowledgements The authors thank numerous members of staff at Kruger National Park for assistance. In particular, we are grateful to Michele Hofmeyr, Bruce Leslie, Harry Biggs (Kruger National Park) and Bob Scholes (CSIR) for constructive advice and assistance; Andre Potgeiter (Kruger National Park) for planting the tree at Skukuza!; Molly McKnight (Oregon) and Million Cossa (Kruger National Park) for help with fieldwork; Gill Collet (CSIR) and Paula Stanlake (University of Wales Swansea) for technical support; and John Vogel (CSIR) and Glenda Swart (South African Weather Service) for the provision of unpublished data. Permission to carry out research was granted by the Conservation Services Management Committee at Kruger National Park. The authors (IR, CAF and NJL) thank the European Union (grants EVK2-CT-2002-00136 and EVK2-CT-2002-00147) and NERC (NE/B501504/ 1) for financial assistance. References Baum, D.A., 1995. A systematic revision of Adansonia (Bombacaceae). Ann. Mo. Bot. Garden 82, 440–471. Coates Palgrave, M., Coates Palgrave, K., 2002. Trees of Southern Africa, third ed. Struik Publishers (Pty) Ltd., Cape Town. Coplen, T., 1995. The discontinuance of SMOW and PDB. Nature 375, 285. Couralet, C., Saas-Klaassen, U., Sterck, F., Bekele, T., Zuidema, P.A., 2005. Combining dendrochronology and matrix modelling in demographic studies: an evaluation for Juniperus procera in Ethiopia. Forest Ecol. Manage. 216, 317–330.
1680
I. Robertson et al. / Applied Geochemistry 21 (2006) 1674–1680
Esterhuyse, N., von Breitenbach, J., So¨hnge, H., 2001. Remarkable Trees of South Africa. Briza Publications, Pretoria. Farquhar, G.D., O’Leary, M.H., Berry, J.A., 1982. On the relationship between carbon isotope discrimination and the intercellular carbon dioxide concentration in leaves. Aust. J. Plant Physiol. 9, 121–137. Fenner, M., 1980. Some measurements on the water relations of baobab trees. Biotropica 12, 205–209. Guy, G.L., 1970. Adansonia digitata and its rate of growth in relation to rainfall in south Central AfricaProceed Transaction, vol. 54. Rhodesia Scientific Association, Salisbury, pp. 68–84. Guy, G.L., 1982. Baobabs and elephants. Afr. J. Ecol. 20, 215– 220. Hua, .Q., Barbetti, M., Worbes, M., Head, J., Levchenko, V.A., 1999. Review of radiocarbon data from atmospheric and tree ring samples for the period 1945–1997 AD. IAWA J. 20, 261– 283. Hua, Q., Barbetti, M., Zoppi, U., Chapman, D.M., Thomson, B., 2003. Bomb radiocarbon calibration from northern New South Wales, Australia: implications for dendrochronology, atmospheric transport, and air-sea exchange of CO2. Radiocarbon 45, 431–447. Kelly, H.L.P., 2000. The effect of elephant utilisation of the Sterculia rogersii and Adansonia digitata populations of the Kruger National Park. Unpublished MSc Thesis, University of Pretoria. LaMarche Jr., V.C., Holmes, R.L., Dunwiddie, P.W., Drew, L.G., 1979. Tree-Ring Chronologies of the Southern Hemisphere: 5. South Africa. Chronology Series V, Laboratory of Tree-Ring Research, University of Arizona. Livingstone, D., 1861. A Popular Account of Missionary Travels and Researches in South Africa. John Murray, London. McCarroll, D., Loader, N.J., 2004. Stable isotopes in tree rings. Quatern. Sci. Rev. 23, 771–801. Robertson, I., Woodborne, S.W., 2002. Carbon isotopes confirm the competitive advantages of Prosopis over Acacia erioloba.
In: Study of Environmental Change using Isotope Techniques, IAEA-CSP-13/P. International Atomic Energy Agency, Vienna, pp. 33–434. Schleser, G.H., Helle, G., Lu¨cke, A., Vos, H., 1999. Isotope signals as climate proxies: the role of transfer functions in the study of terrestrial archives. Quatern. Sci. Rev. 18, 927–943. Schweingruber, F.H., 1988. Tree Rings: Basics and Applications of Dendrochronology. D. Reidel, Boston. Stahle, D.W., Mushove, P.T., Cleveland, M.K., Roig, F., Haynes, G.A., 1999. Management implications of annual growth rings in Pterocarpus angolensis from Zimbabwe. Forest Ecol. Manage 124, 217–229. Stokes, M.A., Smiley, T.L., 1963. An Introduction to Tree-Ring Dating. University of Chicago Press, Chicago. Swart, E.R., 1963. The age of the baobab tree. Nature 373, 134– 137. Tans, P.P., Mook, W.G., 1980. Past atmospheric CO2 levels and 13 C/12C ratios in tree rings. Tellus 32, 268–283. van Wyk, P., 2000. Field Guide to the Trees of the Kruger National Park. Struik Publishers (Pty) Ltd., Cape Town. Vogel, J.C., 1980. Fractionation of the carbon isotopes during photosynthesis. In: Sitzungsberichte der Heidelberger Akademie der Wissenschaften. Springer-Verlag, Berlin, pp. 111–135. Vogel, J.C., Marais, M., 1971. Pretoria radiocarbon dates I. Radiocarbon 13, 378–394. Vogel, J.C., Fuls, A., Visser, A., 2001. Radiocarbon adjustments to the dendrochronology of a yellowwood tree. S. Afr. J. Sci. 97, 164–166. Vogel, J.C., Fuls, A., Visser, A., 2002. Accurate dating with radiocarbon from the atom bombtests. S. Afr. J. Sci. 98, 437– 438. Weyerhaeuser, F.J., 1985. Survey of elephant damage to baobabs in Tanzania’s Lake Manyara National Park. Afr. J. Ecol. 23, 235–243. Wickens, G.E., 1982. The Baobab Africa’s upside down tree. Kew Bull. 37, 173–209. Wilson, R.T., 1988. Vital statistics of the baobab (Adansonia digitata). Afr. J. Ecol. 26, 197–206.